According to the National Institutes of Health (NIH), the United States experienced a staggering 104% increase in analgesic drug prescriptions, such as opioids, over recent years. This surge is part of a broader trend that has had significant public health implications, contributing to a rise in substance abuse and related fatalities.
This project focuses on analyzing drug- and alcohol-induced mortality rates in the United States from 1999 to 2022, emphasizing race and gender as key variables. These demographic dimensions are closely tied to structural inequalities, cultural contexts, and access to resources, all of which can profoundly influence health outcomes. By understanding these connections, we can better appreciate how different populations are affected by substance use disorders and the subsequent mortality rates.
Our variable of interest, Crude Rate, provides a standardized measure of mortality rates per 100,000 individuals, enabling direct comparisons across different races and genders. This rate helps normalize the data, making it easier to observe and analyze trends and disparities.
This study specifically seeks to answer several critical questions: - How do drug and alcohol-induced mortality rates differ across racial groups and genders? - What trends can be observed over time in these mortality rates? - Which demographic groups are most at risk?
Through this comprehensive analysis, we aim to shed light on the complex interplay of race, gender, and substance-related mortality, ultimately contributing to more informed and equitable public health solutions.
This study aims to explore drug- and alcohol-induced mortality rates, focusing on the explanatory variable Crude Rate and its relationship to key demographic dimensions: Race, Gender, State overtime. Below, we outline the data preparation and analytical approaches used to achieve this goal.
The combined datasets, spanning 1999–2020 and 2021–2022, provide a comprehensive foundation for analysis. These datasets include demographic information such as race, gender, state, cause of death, and crude rates. To ensure consistency and relevance in the analysis, the following preprocessing steps were undertaken:
Filtering Unreliable Data
Rows with crude rates labeled as “Unreliable” or “Suppressed” were
removed to maintain data accuracy and reliability.
Merging Race Categories
Removing Unnecessary Columns
Columns not relevant to the analysis, such as notes and redundant
identifiers, were dropped to streamline the dataset.
Handling Missing Values
Missing values were removed using the na.omit() function to
ensure that all analyses were conducted on complete cases, avoiding
potential biases from imputation.
The primary variable of interest is the Crude Rate,
a standardized measure of drug- and alcohol-induced mortality rates per
100,000 people. The demographic dimensions include:
- Race: Adjusted categories include Asian (inclusive of
Native Hawaiian and Pacific Islander), Black or African American, White,
and American Indian or Alaska Native.
- Gender: Male and Female.
- State: Geographic representation within the United
States.
- Year: Temporal range from 1999 to 2022.
The analysis employed the following methods to explore patterns and
disparities in the data:
- Group-Wise Summaries: Mortality rates were summarized
by race, gender, and year to identify patterns and disparities.
- Trends Over Time: Temporal trends in crude rates were
analyzed to understand changes across demographic groups.
- Comparative Analysis: Differences between racial and
gender groups were examined to highlight disparities.
These techniques were selected to ensure a robust exploration of the relationships between demographic factors and mortality rates.
Dynamic figures were created within this R Markdown file using R’s
ggplot2 and plotly packages. These
visualizations include:
- Grouped bar charts to compare crude rates across races and
genders.
- Line charts to illustrate temporal trends in mortality rates.
- Box plots to summarize crude rate distributions by race and
gender.
Suppressed data points, typically used to protect confidentiality when numbers are too small, were excluded from the analysis to ensure accuracy and consistency. By focusing on reliable and complete data, this approach minimizes potential misrepresentation.
Through this structured preprocessing and analytical approach, we aim to uncover meaningful insights into the disparities and trends in drug- and alcohol-induced mortality rates.
First, as previously mentioned, to ensure the integrity of our
findings, we started by evaluating whether the cleaning
process—specifically, the removal of unreliable and suppressed
data—introduced any significant bias into our analysis. This initial
analysis compares the distributions of the primary explanatory variable,
Crude Rate (drug and alcohol-induced deaths), between the
cleaned dataset (combined_data.df) and the suppressed
dataset (combined_sup.df).
Methods
Two complementary approaches were employed to assess the impact of data cleaning: 1. Kolmogorov-Smirnov (KS) Test: - A non-parametric statistical test was used to compare the distributions of Crude Rate in the cleaned and suppressed datasets. - The null hypothesis (H₀) states that the distributions of Crude Rate in both datasets are identical. 2. Visual Comparison: - Density plots were generated to provide a visual assessment of the distributions.
Kolmogorov-Smirnov Test -
D-statistic: 0.010988
- p-value: 0.2278
- Conclusion: The p-value exceeds the common
significance level of 0.05, indicating no statistically significant
difference between the two distributions. Thus, we fail to reject the
null hypothesis, confirming that the distributions of Crude
Rate are not significantly affected by the data cleaning
process.
** Visual Comparison**
Discussion
The results from both the KS test and density plot corroborate that the data cleaning process did not introduce significant biases or alter the overall distribution of Crude Rate. This finding is critical, as it validates the integrity of the cleaned dataset for subsequent analyses. The removal of suppressed and unreliable data ensures analytical clarity while preserving the representatives of the trends. By demonstrating that the cleaned dataset is statistically comparable to the suppressed dataset, we can proceed with confidence in exploring the relationships between Crude Rate and race and gender.
This analysis highlights notable demographic disparities in drug and alcohol-related deaths:
For this section, we analyzed the average crude rates of drug and alcohol-related deaths across racial groups, genders, using causes of death. The dataset spans the years 1999 to 2022, and the plot is faceted by race, with color used to represent specific causes and shape used to denote gender.
Temporal patterns are displayed along the x-axis, while crude rates are shown on the y-axis.
The visualization was generated using ggplot2 and
plotly, allowing for an interactive exploration of the
data. The code first groups and summarizes the data by year, race,
gender, and cause of death. Average crude rates were calculated for each
subgroup, excluding missing values. A viridis color scale
was used to enhance visual clarity, and the interactive nature of the
plot enables deeper exploration of trends across these dimensions.
Persistent Disparities by Race - Native populations exhibit consistently higher rates of alcohol- and drug-related deaths compared to other racial groups. This finding aligns with existing literature documenting significant health inequities in these communities.
Gender-Specific Patterns Gender differences remain
consistent to the previous plots. This version does highlight the higher
number of Males dying of
All other alcohol-induced causes.
Cause-Specific Trends The top three causes of death
were“All other alcohol-induced causes”, “All other drug-induced
causes”, and “Drug poisonings (overdose), unintentional”.
Causes such as Drug Poisoning (Homicide) and
Drug Poisoning (Suicide) had comparatively lower rates.
Temporal Trends Over the study period (1999–2022), fluctuations in crude rates reveal important temporal patterns: - Certain years show significant spikes in crude rates, potentially linked to societal events or public health crises (e.g., peaks in opioid-related deaths). - The most recent years suggest a plateau or slight decline in some racial groups, reflecting potential improvements in public health interventions or data collection.
This analysis underscores the ongoing public health challenge of drug- and alcohol-related mortality in the United States. The findings highlight critical racial and gender disparities, with Native populations disproportionately affected and males experiencing consistently higher rates than females. These disparities reflect deep-seated structural inequities and emphasize the need for targeted interventions.
To build on this analysis, future studies should:
In conclusion, the analysis of drug and alcohol-related deaths from 1999 to 2022 highlights persistent disparities in mortality rates across racial groups, genders. It also displays which causes of death within the dataset are most prevalent. Native populations, in particular, exhibit disproportionately high rates, emphasizing the need for culturally tailored prevention and intervention strategies. Further, Males exhibited higher crude rates than their Female counterparts. Additionally, the overall leading causes of death were unintentional drug poisonings, alcohol-related causes, and other drug-induced conditions—remain key areas for public health action.
The findings reinforce the importance of addressing the root causes of these disparities, including systemic inequities, access to healthcare, and cultural barriers to treatment. By illuminating these patterns, this analysis provides a foundation for informed public health strategies aimed at reducing preventable deaths and improving health equity nationwide.
References:
https://pmc.ncbi.nlm.nih.gov/articles/PMC3955827/#:~:text=The%20estimated%20total%20number%20of,11.2%E2%80%9312.4
https://www.ncbi.nlm.nih.gov/books/NBK458661/
https://americanaddictioncenters.org/addiction-statistics/native-americans